Re: PANIC: can't juggle and compress too

Freebsd 4.9
Postgres 7.4Beta5

We have a daemon that among other things vacuums the channels table
every 5 minutes. Worked fine until last night.

Oct 28 01:04:49 ... postgres[7572]: [52226-1] LOG: statement: VACUUM ANALYZE channels
Oct 28 01:04:50 ... postgres[7572]: [52227-1] PANIC: can't juggle and compress too
Oct 28 01:04:50 ... postgres[7572]: [52227-2] STATEMENT: VACUUM ANALYZE channels
Oct 28 01:04:50 ... postgres[7565]: [1-1] LOG: server process (PID 7572) was terminated by signal 6
Oct 28 01:04:50 ... postgres[7565]: [2-1] LOG: terminating any other active server processes
Oct 28 01:04:50 ... postgres[7630]: [7841-1] WARNING: terminating connection because of crash of another server process
Oct 28 01:04:50 ... postgres[7630]: [7841-2] DETAIL: The postmaster has commanded this server process to roll back the current transaction and exit, because ano ther server
Oct 28 01:04:50 ... postgres[7576]: [47833-1] WARNING: terminating connection because of crash of another server process
Oct 28 01:04:50 ... postgres[7630]: [7841-3] process exited abnormally and possibly corrupted shared memory.
Oct 28 01:04:50 ... postgres[7577]: [6-1] WARNING: terminating connection because of crash of another server process
...

I googled can't juggle and compress too and the only link is to the
actual source.

Bug?

Regards,

Dror

--
Dror Matalon
Zapatec Inc
1700 MLK Way
Berkeley, CA 94709
http://www.zapatec.com

Dror Matalon <dror(at)zapatec(dot)com> writes:
> Oct 28 01:04:50 ... postgres[7572]: [52227-1] PANIC: can't juggle and compress too

Oh? I don't suppose you have a way to reproduce this :-(

regards, tom lane

On Tue, Oct 28, 2003 at 02:55:17PM -0500, Tom Lane wrote:
> Dror Matalon <dror(at)zapatec(dot)com> writes:
> > Oct 28 01:04:50 ... postgres[7572]: [52227-1] PANIC: can't juggle and compress too
>
> Oh? I don't suppose you have a way to reproduce this :-(

No not really. It's been running for a few hours without any problems.
Looks like after this happened the database shut down all processes and
then did a "recovery."

What does the error message mean?

Regards,

Dror
>
> regards, tom lane
>
> ---------------------------(end of broadcast)---------------------------
> TIP 6: Have you searched our list archives?
>
> http://archives.postgresql.org

--
Dror Matalon
Zapatec Inc
1700 MLK Way
Berkeley, CA 94709
http://www.zapatec.com

Dror Matalon <dror(at)zapatec(dot)com> writes:
> On Tue, Oct 28, 2003 at 02:55:17PM -0500, Tom Lane wrote:
>> Dror Matalon <dror(at)zapatec(dot)com> writes:
> Oct 28 01:04:50 ... postgres[7572]: [52227-1] PANIC: can't juggle and compress too

> What does the error message mean?

It's an unimplemented case in FSM memory management. I thought that
it was a can't-happen case; either I was wrong and it does need to be
implemented, or there's a bug elsewhere that's triggering the failure
condition. It'd be a lot easier to figure out what's wrong if I could
reproduce the failure ...

regards, tom lane

Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> writes:
> Dror Matalon <dror(at)zapatec(dot)com> writes:
>> Oct 28 01:04:50 ... postgres[7572]: [52227-1] PANIC: can't juggle and compress too

>> What does the error message mean?

> It's an unimplemented case in FSM memory management. I thought that
> it was a can't-happen case; either I was wrong and it does need to be
> implemented, or there's a bug elsewhere that's triggering the failure
> condition.

I've concluded that that case does need to be handled; I think it can
arise in situations where a relation's allocated share of FSM space
decreases due to roundoff error. Fix committed accordingly.

regards, tom lane

On Wed, Oct 29, 2003 at 12:39:57PM -0500, Tom Lane wrote:
> Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> writes:
> > Dror Matalon <dror(at)zapatec(dot)com> writes:
> >> Oct 28 01:04:50 ... postgres[7572]: [52227-1] PANIC: can't juggle and compress too
>
> >> What does the error message mean?
>
> > It's an unimplemented case in FSM memory management. I thought that
> > it was a can't-happen case; either I was wrong and it does need to be
> > implemented, or there's a bug elsewhere that's triggering the failure
> > condition.
>
> I've concluded that that case does need to be handled; I think it can
> arise in situations where a relation's allocated share of FSM space
> decreases due to roundoff error. Fix committed accordingly.

The best support money can't buy. One more reason to love Postgres.

Thanks Tom,

Dror

>
> regards, tom lane

--
Dror Matalon
Zapatec Inc
1700 MLK Way
Berkeley, CA 94709
http://www.zapatec.com

Tom Lane wrote:
> Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> writes:
>
>>Dror Matalon <dror(at)zapatec(dot)com> writes:
>>
>>>Oct 28 01:04:50 ... postgres[7572]: [52227-1] PANIC: can't juggle and compress too
>
>
>>>What does the error message mean?
>
>
>>It's an unimplemented case in FSM memory management. I thought that
>>it was a can't-happen case; either I was wrong and it does need to be
>>implemented, or there's a bug elsewhere that's triggering the failure
>>condition.
>
>
> I've concluded that that case does need to be handled; I think it can
> arise in situations where a relation's allocated share of FSM space
> decreases due to roundoff error. Fix committed accordingly.

Less dangerous then this:

http://mathworld.wolfram.com/RoundoffError.html

:-O

Regards
Gaetano Mendola